Spark Structured Streaming框架(3)之数据输出源详解

　　Spark Structured streaming API支持的输出源有：Console、Memory、File和Foreach。其中Console在前两篇博文中已有详述，而Memory使用非常简单。本文着重介绍File和Foreach两种方式，并介绍如何在源码基本扩展新的输出方式。

1. File

　　Structured Streaming支持将数据以File形式保存起来，其中支持的文件格式有四种:json、text、csv和parquet。其使用方式也非常简单只需设置checkpointLocation和path即可。checkpointLocation是检查点保存的路径，而path是真实数据保存的路径。

如下所示的测试例子：

// Create DataFrame representing the stream of input lines from connection to host:port

val lines = spark.readStream

.format("socket")

.option("host", host)

.option("port", port)

.load()

// Split the lines into words

val words = lines.as[String].flatMap(_.split(" "))

// Generate running word count

val wordCounts = words.groupBy("value").count()

// Start running the query that prints the running counts to the console

val query = wordCounts.writeStream

.format("json")

.option("checkpointLocation","root/jar")

.option("path","/root/jar")

.start()

注意：

File形式不能设置"compelete"模型，只能设置"Append"模型。由于Append模型不能有聚合操作，所以将数据保存到外部File时，不能有聚合操作。

2. Foreach

　　foreach输出方式只需要实现ForeachWriter抽象类，并实现三个方法，当Structured Streaming接收到数据就会执行其三个方法，如下的测试示例：

/*

* Licensed to the Apache Software Foundation (ASF) under one or more

* contributor license agreements. See the NOTICE file distributed with

* this work for additional information regarding copyright ownership.

* The ASF licenses this file to You under the Apache License, Version 2.0

* (the "License"); you may not use this file except in compliance with

* the License. You may obtain a copy of the License at

*

* http://www.apache.org/licenses/LICENSE-2.0

*

* Unless required by applicable law or agreed to in writing, software

* distributed under the License is distributed on an "AS IS" BASIS,

* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.

* See the License for the specific language governing permissions and

* limitations under the License.

*/

// scalastyle:off println

package org.apache.spark.examples.sql.streaming

import org.apache.spark.sql.SparkSession

/**

* Counts words in UTF8 encoded, '\n' delimited text received from the network.

*

* Usage: StructuredNetworkWordCount <hostname> <port>

* <hostname> and <port> describe the TCP server that Structured Streaming

* would connect to receive data.

*

* To run this on your local machine, you need to first run a Netcat server

* `$ nc -lk 9999`

* and then run the example

* `$ bin/run-example sql.streaming.StructuredNetworkWordCount

* localhost 9999`

*/

object StructuredNetworkWordCount {

def main(args: Array[String]) {

if (args.length < 2) {

System.err.println("Usage: StructuredNetworkWordCount <hostname> <port>")

System.exit(1)

}

val host = args(0)

val port = args(1).toInt

val spark = SparkSession

.builder

.appName("StructuredNetworkWordCount")

.getOrCreate()

import spark.implicits._

// Create DataFrame representing the stream of input lines from connection to host:port

val lines = spark.readStream

.format("socket")

.option("host", host)

.option("port", port)

.load()

// Start running the query that prints the running counts to the console

val query = wordCounts.writeStream

.outputMode("append")

.foreach(new ForearchWriter[Row]{

override def open(partitionId:Long,version:Long):Boolean={

println("open")

return true

}

override def process(value:Row):Unit={

val spark = SparkSession.builder.getOrCreate()

val seq = value.mkString.split(" ")

val row = Row.fromSeq(seq)

val rowRDD:RDD[Row] = sparkContext.getOrCreate().parallelize[Row](Seq(row))

val userSchema = new StructType().add("name","String").add("age","String")

val peopleDF = spark.createDataFrame(rowRDD,userSchema)

peopleDF.createOrReplaceTempView(myTable)

spark.sql("select * from myTable").show()

}

override def close(errorOrNull:Throwable):Unit={

println("close")

}

})

.start()

query.awaitTermination()

}

}

// scalastyle:on println

　　上述程序是直接继承ForeachWriter类的接口，并实现了open()、process()、close()三个方法。若采用显示定义一个类来实现，需要注意Scala的泛型设计，如下所示：

class myForeachWriter[T<:Row](stream:CatalogTable) extends ForearchWriter[T]{

override def open(partionId:Long,version:Long):Boolean={

println("open")

true

}

override def process(value:T):Unit={

println(value)

}

override def close(errorOrNull:Throwable):Unit={

println("close")

}

}

3. 自定义

　　若上述Spark Structured Streaming API提供的数据输出源仍不能满足要求，那么还有一种方法可以使用：修改源码。

如下通过实现一种自定义的Console来介绍这种使用方式：

3.1 ConsoleSink

　　Spark有一个Sink接口，用户可以实现该接口的addBatch方法，其中的data参数是接收的数据，如下所示直接将其输出到控制台：

class ConsoleSink(streamName:String) extends Sink{

override def addBatch(batchId:Long, data;DataFrame):Unit = {

data.show()

}

}

3.2 DataStreamWriter

　　在用户自定义的输出形式时，并调用start()方法后，Spark框架会去调用DataStreamWriter类的start()方法。所以用户可以直接在该方法中添加自定义的输出方式，如我们向其传递上述创建的ConsoleSink类示例，如下所示：

def start():StreamingQuery={

if(source == "memory"){

...

}else if(source=="foreach"){

...

}else if(source=="consoleSink"){

val streamName:String = extraOption.get("streamName") mathc{

case Some(str):str

case None=>throw new AnalysisException("streamName option must be specified for Sink")

}

val sink = new consoleSink(streamName)

df.sparkSession.sessionState.streamingQueryManager.startQuery(

extraOption.get("queryName"),

extraOption.get("checkpointLocation"),

df,

sink,

outputMode,

useTempCheckpointLocaltion = true,

recoverFromCheckpointLocation = false,

trigger = trigger

)

}else{

...

}

}

3.3 Structured Streaming

　　在前两部修改和实现完成后，用户就可以按正常的Structured Streaming API方式使用了，唯一不同的是在输出形式传递的参数是"consoleSink"字符串，如下所示：

def execute(stream:CatalogTable):Unit={

val spark = SparkSession

.builder

.appName("StructuredNetworkWordCount")

.getOrCreate()

/**1. 获取数据对象DataFrame*/

val lines = spark.readStream

.format("socket")

.option("host", "localhost")

.option("port", 9999)

.load()

/**2. 启动Streaming开始接受数据源的信息*/

val query:StreamingQuery = lines.writeStream

.outputMode("append")

.format("consoleSink")

.option("streamName","myStream")

.start()

query.awaitTermination()

}

4. 参考文献

[1]. Structured Streaming Programming Guide.

[2]. Kafka Integration Guide.